Example pangenome hash correlation: Ralstonia solanacearum
===

## Setup

mamba create -n smash-pg sourmash=4.8.10
mamba activate smash-pg
pip install sourmash_utils sourmash_plugin_pangenomics sourmash_plugin_betterplot

git clone [e-mail]:ctb/2024-pangenome-hash-corr.git
cd 2024-pangenome-hash-corr
mkdir rs-pangenome
cd rs-pangenome

> if the git@ link doesn't work for you, use `https://github.com/ctb/2024-pangenome-hash-corr.git`

## Download sketches and taxonomic information

Download a database containing signatures of 32 Ralstonia genomes (pathogenic and not) and the corresponding taxonomic and lingroup information.

# database
curl -JLO https://osf.io/download/wxtk3/

# taxonomy csv
curl -JLO https://osf.io/download/sj2z7/

# lingroup csv
curl -JLO https://osf.io/download/nqms2/

# phylotype categories
curl -JLO https://osf.io/download/ptvgr/

## Build merged signature
Merge k-mers from all Ralstonia genomes into a single sourmash signature representing the "pangenome".

sourmash scripts pangenome_merge -k 31\
    ralstonia.sc1000.zip \
    -o ralstonia32.k31-sc1000.merged.zip

## Build ranktable, no lineages
Build a table of the k-mers in the ralstonia pangenome

sourmash scripts pangenome_ranktable \
    ralstonia32.k31-sc1000.merged.zip \
    -o ralstonia32.k31-sc1000.ranktable.csv \
    -k 31 

## Calculate hash presence
calculate presence of each k-mer in each ralstonia genome.

../calc-hash-presence.py ralstonia32.k31-sc1000.ranktable.csv \
    ralstonia.sc1000.zip -o ralstonia32.k31-sc1000.dump \
    --scaled=1000

## Rectangular matrix showing hash x genome correlations:
Build a presence-absence table from the hash presence counted above. Limit the total number of k-mers so we can plot in the next step (here, we limit to k-mers found in 15 or more genomes).

../hash-by-sample.py ralstonia32.k31-sc1000.dump \
    -o ralstonia32.k31-sc1000.presence.csv -m 15 

output:

loaded 42616 hash to sample entries.
wrote 42616 entries to 'ralstonia32.k31.presence.csv'
use 'sourmash scripts clustermap1' from betterplot to plot!
e.g. 'sourmash scripts clustermap1 ralstonia32.k31.presence.csv -o fig.png

## Try to plot
cluster and plot k-mers per genome. Add in the known phylogroup structure.

sourmash scripts clustermap1 ralstonia32.k31-sc1000.presence.csv -u presence \
    -o ralstonia32.k31-sc1000.presence.png --boolean  -R ralstonia32.phylotypes.csv --figsize-x 13 --figsize-y 10

> labels are not that useful. Try adding `--no-lables` for a prettier plot